Characterization of Mutations in NOT2 Indicates that it Plays an Important Role in Maintaining the...

Characterization of Mutations in NOT2 Indicates that itPlays an Important Role in Maintaining the Integrity ofthe CCR4–NOT Complex

Pamela Russell1, John D. Benson2 and Clyde L. Denis1*

1Department of Biochemistryand Molecular BiologyUniversity of New Hampshire46 College Road, Durham, NH03824, USA

2ENANTA Pharmaceuticals500 Arsenal Street, WatertownMA 02472, USA

The NOT2 protein is a component of the CCR4–NOT complex that playsmultiple roles in the regulation of mRNA production in the yeastSaccharomyces cerevisiae. We have identified four novel not2 mutationsand have characterized these and two previously described alleles as tothe means by which they affect CCR4–NOT function. While two of thenot2 alleles, not2-4 (carrying a G31R alteration) and not2<L9P, resulted insevere growth defects and caused a not phenotype at the HIS3 locus,these phenotypes appear to arise from partially different effects. Thenot2<L9P mutation resulted in complete loss of the 1.9 £ 106 Da(1.9 MDa) CCR4–NOT complex, and the not2<L9P protein displayedincreased ability to associate with the NOT5 protein. In contrast, thenot2-4 allele destabilized the CCR4–NOT complex to a lesser extent andhad no effect on NOT5 association with NOT2. Instead, as previouslyreported, it displayed defective interactions with ADA2, a component ofthe SAGA complex. The not2<R165G also abrogated NOT2 ability tointeract with ADA2 but had little effect on the integrity of the CCR4–NOT complex. Alterations in NOT2 contacts to ADA2, therefore, do notnecessarily result in effects on the CCR4–NOT complex nor result insevere growth defects. We also observed that the four NOT2 N-terminalmutations affected NOT5 association with the CCR4–NOT complexes,suggesting that it is the N terminus of NOT2 that contacts and stabilizesNOT5 interactions. These results indicate that it is the loss of the integrityof the CCR4–NOT complex which leads to severe not2 phenotypes andthat the NOT2 contacts to ADA2 play a lesser role in NOT2 function.

q 2002 Elsevier Science Ltd. All rights reserved

Keywords: transcription; mRNA degradation; CCR4–complex; SAGA;yeast*Corresponding author

Introduction

One of the important regulatory complexesinvolved in eukaryotic gene transcription is theCCR4–NOT complex.1 This unique complex hasboth positive and negative affects on geneexpression and regulates genes involved in cellcycle control, non-fermentative processes, and cellwall integrity.2 – 5 Although it has been shown thatthe CCR4–NOT complex is distinct from otherlarge protein complexes such as the SAGA/ADAcomplex, the SWI/SNF complex, the RNA poly-merase II holoenzyme, and TFIID,1,6,7 it remains

clear that components of the CCR4–NOT complexinteract with members of these large protein com-plexes to subsequently affect transcriptionalregulation.8 – 11

The CCR4–NOT complex exists in two forms,1.0 £ 106 Da (1.0 MDa) and 1.9 MDa in size.1,12,13

Gel filtration chromatography and mass spec-trometry analysis have verified the presence of theCCR4, CAF1, CAF40, CAF130, NOT1, NOT2,NOT3, NOT4 and NOT5 proteins in the 1.0 MDacomplex.1,12,13 Of these factors only NOT1 is essen-tial. The larger of the CCR4–NOT complexes isthought to contain the components of the 1.0 MDacomplex and possibly several other proteins.5,11,14,15

The physical and functional separation of theCCR4 and CAF1 proteins from the NOT2, NOT4and NOT5 proteins in the CCR4–NOT complexsuggests that these two subsets of proteins perform

0022-2836/02/$ - see front matter q 2002 Elsevier Science Ltd. All rights reserved

E-mail address of the corresponding author:[email protected]

Abbreviations used: 3-AT, 3-aminotriazole; 6-AU,6-azauracil; GST, glutathione-S-transferase.

doi:10.1016/S0022-2836(02)00707-6 available online at http://www.idealibrary.com onBw

J. Mol. Biol. (2002) 322, 27–39

separate functions.12 For example, recent studiesindicate that the CCR4 and CAF1 protein areinvolved directly in mRNA degradation in thecytoplasm with the NOT proteins having lessereffects.16 – 19 In contrast, the NOT genes werenamed for negative on TATA,20,21 indicating theirability to negatively regulate transcription.Mutations in the five NOT genes cause increasedtranscription in many chromosomal and reportergenes,1,22 – 24 although some positive effects on tran-scription have also been observed.1 It has beensuggested that TFIID is a target of the NOT pro-teins, particularly NOT5, and that the sequesteringof TFIID by the CCR4–NOT complex may restrictthe expression of certain genes.8,10,23 Moreover,human NOT4 has been found to contain a ubiqui-tin protein ligase activity suggestive of a role inprotein degradation.25

In an effort to further characterize the coactivatorfunction of the CCR4–NOT complex, we havefocused on the role of the NOT2 protein in theCCR4–NOT complex. NOT2 has been shown to bemost closely associated with NOT5.12 Recent find-ings suggested that the relatively small 26 kDaNOT2 protein contained two functional regions.9

The C-terminal domain was thought to be involvedin the evolutionarily conserved function of theCCR4–NOT complex, and the N-terminal domainof NOT2 involved in contacting ADA2.9 This con-tact to ADA2 also correlated with the ability ofLexA-NOT2 to activate transcription.

We have assessed six individual point mutationsin the NOT2 protein using genetic and biochemicaltechniques in order to dissect the functions ofNOT2. These mutations were categorized by theireffects on the structural integrity of the CCR4–NOT complex and by their ability to physicallyand functionally alter NOT2 interactions withADA2 that lies outside of the CCR4–NOT com-plex. Our results indicate that NOT2 plays animportant role in maintaining the integrity of theCCR4–NOT complex and that not2 defects are notsimply the result from effects in contacting ADA2.

Results

Identifying mutations in NOT2 that alter itstranscriptional activation function

Previous reports indicated that NOT2, whenexpressed as a LexA fusion protein, is a strongtranscriptional activator21 and that this activity isabrogated by the not2-4 mutation (contains theG31R alteration).9 However, in the yeast strainMaV103, commonly used for the yeast two-hybridscreen,26 the GAL4 DNA binding domain (GAL4)fused to the NOT2 protein does not activate theGAL4 responsive reporters lacZ, HIS3 and URA3(data not shown, Table 1). In contrast, GAL4-NOT2-N (expressing amino acid residues 1–102 ofNOT2) is a strong activator in this system. Thisactivity is lost in GAL4-not2-4-N when the not2-4

allele is expressed (Table 1). This observationsuggests that the C terminus of NOT2 may blockthe transcriptional activation function of the NOT2N terminus and that mutant NOT2 alleles couldbe identified that could evade the C-terminalinhibition of that activity.

To create mutant not2 alleles, wild-type NOT2was used as a template to perform PCR-muta-genesis. PCR fragments, containing randommutations in the NOT2 gene, were cloned into theGAL4 expression plasmid, and GAL4–NOT2fusions were tested in MaV103 for activation ofthe HIS3 reporter (Table 1). 3-Aminotriazole(3-AT), a competitive inhibitor of the HIS3 geneproduct,27 was used in this screen, and positivetransformants were verified through their corre-sponding activation of the URA3 and lacZreporters. Plasmids containing the mutant not2alleles were rescued and their phenotypes wereconfirmed in freshly transformed naive MaV103.Plasmids conferring confirmed phenotypes weresequenced and their NOT2 coding sequences weresubcloned into the pRS316 expression vector.Table 2 describes the mutant not2 alleles identifiedin this screen. In addition to these new mutations

Table 1. Selection of the not2 mutants using a HIS3reporter

GAL4-NOT2 expression plasmid 3-AT (10 mM)

GAL4-NOT2 2GAL4-NOT2-N þGAL4-not2-4-N 2GAL4-not2<L9P þGAL4-not2<P59S þGAL4-not2<W60C þGAL4-not2<K134E þGAL4-not2<R165G þ

PCR-derived not2 fragments were cloned into a GAL4expression plasmid. MaV103 was then transformed with theGAL4-NOT2 DNA and incubated on medium containing10 mM 3-AT, an inhibitor of the HIS3 gene. Growth of MaV103on this medium indicates activation of the HIS3 reporter geneby the mutated GAL4–not2 fusion protein. GAL4-NOT2-N andGAL4-not2-4-N refer to the N-terminal portion of NOT2 ornot2-4 fused to GAL4.

Table 2. Mutations in the NOT2 coding sequence

Allele Mutation Codon change Publication

not2-1 (MY16) G3C M1Stop 9,21not2-4 (MY22) G91C G31R 9,21not2-3 G180T W60C This worknot2-5 T26C L9P This worknot2-6 A493G R165G This worknot2-7 A400G K134E This worknot2-10 C175T P59S 28; this work

The K18R substitution originally identified for not2-49 doesnot appear to be either part of the original not2-4 allele or to con-tribute to the not2-4 allele effect on LexA-not2-4 function (datanot shown). All not2 alleles except not2-1 were expressed inyeast on a pRS316 plasmid.

28 NOT2 is Required for CCR4–NOT Complex Integrity

and not2-4, we also sequenced the not2-10 allelepreviously described.28 This not2 allele was identi-fied in an independent screen for positive andnegative regulators of the RNA polymerase IIholoenzyme.28 The mutant not2-10 allele,not2<P59S, was identified as capable of suppres-

sing the temperature sensitive phenotype con-ferred by rpb2-3, a mutation in the second largestsubunit of RNA polymerase II. The rpb2-3 allele,at the restrictive temperature, drastically reducestranscription from most yeast genes.28 Each ofthese six not2 point mutations characterized here

Figure 1. Sequence comparisons of NOT2 proteins. The NOT2 proteins from Saccharomyces cereviscae (NOT2),Schizosaccharomyces pombe (S. pombe ), human (hNOT2), Drosophila (dNOT2), and Caenorhabditis elegans (cNOT2) werecompared. The sequence at the top represents the most common residues and the bar graph above is proportioned tothe representation of that residue. Residues that are boxed correspond to the alterations presented in Table 2.

NOT2 is Required for CCR4–NOT Complex Integrity 29

were found to be in residues completely conservedamong a range of NOT2 orthologs or to be locatedwithin highly conserved regions of the NOT2protein (Figure 1). Three of the six absolutelyconserved amino acid codons in the N-terminal 60residues of NOT2 were altered by not2<L9P, not2-4 and not2<P59S, suggesting that the mutantscreens had targeted conserved and importantresidues of NOT2.

Phenotypic effects of the not2 alleles

Phenotypic analysis of the NOT2 mutant alleleswas conducted in order to detect altered functionsof the NOT2 protein. Only the not2<L9P alleleresulted in severe phenotypes comparable to thosepreviously associated with the not2-1 and not2-4alleles. The not2<L9P allele resulted in 3-ATresistance gene (indicative of increased HIS3expression) in a strain lacking a functional GCN4gene,21 displayed sensitivity to 6AU (suggestive ofa defect in elongation),29 showed reduced growthat 37 8C under non-fermentative growth con-ditions, and failed to grow at 41 8C under glucosegrowth conditions (Table 3). However, not2<L9Phad less severe caffeine and MgCl2 sensitivityphenotypes than not2-4.

We also analyzed whether the not2 alleles wereable to complement the ability of not2<P59S tosuppress the temperature sensitivity conferred bythe rpb2-3 allele.28 The not2 mutant alleles, not2-4,not2<L9P and not2<W60C were unable to comple-ment the not2<P59S effect (Table 3), whereas wild-type NOT2 and the mutant alleles not2<K134E andnot2<R165G were able to complement not2<P59S.

We also tested for altered transcriptionalactivation ability of the mutant NOT2 alleles byconstructing LexA fusion proteins correspondingto each mutant allele. Each of the LexA–NOT2fusions was expressed in yeast and analyzed forits ability to activate a LexA-lacZ reporter. All of

the LexA–NOT2 variants were expressed at com-parable levels as determined by Western analysis(data not shown). In agreement with the previousreports,9,21 the LexA–NOT2 fusion protein wasable to activate transcription of lacZ (Table 4). Thenot2-4 mutation decreased this activity 20-fold,whereas increased transcriptional activity wasnoted in the LexA fusion proteins containing theNOT2 alterations L9P, P59S, W60C, and K134E,consistent with the original screen that wasused for detecting several of these mutation (seeDiscussion). In contrast, the not2<R165G alleledisplayed a threefold decrease in activation as aLexA fusion protein.

CCR4–NOT complex integrity is affected bymutations in NOT2

The above phenotypic analysis indicates that thenot2-4 and not2<L9P alleles affect a number ofNOT2 functions, whereas several of the other

Table 3. Phenotypes of the mutant not2 alleles

3-AT(10 mM)

6AU(100 mM)

Temperature sensitivity Caff.(15 mM)

MgCl2

(75 mM)Glycerol

(3%) rpb2-3, not2 < P59S30 8C 34 8C 30 8C 37 8C 39 8C 41 8C 30 8C 30 8C 37 8C 37 8C

NOT2 – þ þ þ þ þ þ þ þ –not2-1 þ – þ w – – – – – NDnot2-4 þ – þ þ þ – – – W þnot2<L9P þ – þ þ þ – þ þ W þnot2<P59S – þ þ þ þ þ þ þ þ þnot2<W60C – þ þ þ þ þ þ þ þ þnot2<K134E – þ þ þ þ þ þ þ þ –not2<R165G – þ þ þ þ þ þ þ þ –

Growth was scored on YD plates supplemented with 15 mM caffeine, 75 mM MgCl2 or 3% glycerol as indicated. The strains used inthese screens were isogenic to KY803. 3AT: growth was scored on minimal plates lacking histidine and containing 10 mM 3-amino-triazole (3-AT). 6AU: growth was scored on minimal plates lacking uracil and containing 100 mM 6-azauracil.29 The strains carryingthe rpb2-3 mutation (Z421) (not shown), and the rpb2-3 and not2<P59S mutations (Z1446) were transformed with plasmids containingthe mutant not2 alleles. Failure to complement the not2<P59S allele in a rbp2-3 background resulted in growth at 37 8C. None of theNOT2 alleles allowed growth at 37 8C when expressed in a rpb2-3 NOT2 background. ND, not done; þ , good growth; w, weak growth;2 , no or poor growth.

Table 4. NOT2 alleles affect NOT2 transcriptionalactivation ability

LexA-fusion b-Galactosidase activity

LEXA-NOT2 100 ^ 6.7LEXA-not2-4 4.5 ^ 1.0LEXA-not2<L9P 220 ^ 19LEXA-not2<P59S 290 ^ 34LEXA-not2<W60C 520 ^ 48LEXA-not2<K134E 340 ^ 26LEXA-not2<R165G 35 ^ 3.2

b-Galactosidase assays were conducted after growth of strainEGY188 at 30 8C in minimal media lacking uracil and histidinethat was supplemented with 2% (w/v) glucose. LexA-NOT2fusion proteins contain residues 1–202 of the E. coli LexAprotein, and were expressed in EGY188 containing the lacZreporter, p34 (8 LexA operators). b-Galactosidase activities(units/mg) ^ standard error of the mean represent the averageof at least five separate determinations.


alleles display lesser effects. We subsequentlyanalyzed whether these phenotypic effects werethe result of disturbances in the protein–proteininteractions within the CCR4–NOT complex. Eachof the mutated NOT2 proteins were expressed atendogenous levels in a yeast strain containing thenot2-1 null mutation. All protein variants wereexpressed to comparable levels (data not shown).In order to analyze the integrity of the CCR4–NOT complex, antibodies directed at the CAF1,NOT3 and NOT5 proteins were individuallyincubated with the yeast crude extracts, andimmunoprecipitations were performed.1

The CCR4–NOT complex components CAF1,NOT1, CCR4 and CAF40 were immunoprecipi-tated by the CAF1 antibody in all of the strains,including the not2-1 background strain (Figure 2).This result indicates that NOT2 does not affect theCCR4-CAF1 part of the CCR4–NOT complex inagreement with previous results.12,13 In contrast,the NOT5 and NOT2 proteins were not present inthe CAF1 immunoprecipitates of the strains con-taining the not2-1, not2-4 and not2<L9P alleles.The NOT4 protein was also missing in the CAF1precipitated complex in the strain containing thenull mutant, not2-1. None of the differences wereattributable to diminished levels of the protein inthe crude extracts (data not shown). The crudeextracts of strains containing the not2-1, not2-4 andnot2<L9P alleles, however, had decreased levelsof the NOT3 protein and no NOT3 protein wasdetected in the CAF1-mediated immunoprecipita-tions (data not shown). Clearly, the not2-4 and

Figure 2. The not2-1, not2-4, and not2<L9P allelescause disruption of the CCR4–NOT complex. Immuno-precipitations of crude extracts (2 mg) were conductedwith antibody directed again CAF1. Western analysiswas conducted with antibodies directed against CCR4–NOT protein components as indicated. The strainsutilized contain the respective NOT2 alleles as describedin Table 2. NOT5 and NOT2 proteins were present incrude extracts in the various mutant strains to com-parable extents except for the strain containing the not2-1 allele (see Figure 3).

Figure 3. The NOT2-NOT5 inter-action is stabilized by thenot2<L9P allele. NOT2 and NOT5proteins were identified by Westernanalysis from crude extracts(200 mg) or from NOT5 immuno-precipitates (2 mg of crude extractused). NOT2 alleles are as indicatedin Table 2.


not2<L9P alleles affect NOT2, NOT5, and NOT3association with the CCR4–NOT complex,whereas the other NOT2 alterations had noapparent effect.

The CCR4–NOT complex components were sub-sequently analyzed by immunoprecipitation usingantibody directed at the NOT3 protein. Again, theNOT3 protein levels were diminished in the crude

Figure 4. Gel filtration chromatography of CCR4–NOT complexes. Superose 6 chromatography on crude extractsfrom yeast strains carrying the indicated NOT2 allele was performed as described.1 The markers used for standardiz-ing the column were blue dextran (2.0 MDa), thyroglobulin (0.67 MDa), and bovine serum albumin (0.066 MDa).


Figure 5. NOT2 alleles affect NOT5 protein association in 1.9 MDa complex. Superose 6 gel permeationchromatography was conducted as described in the legend to Figure 4.


extracts of the strains containing the not2-1, not2-4and not2<L9P alleles, resulting in low proteinlevels for all CCR4–NOT components in theimmunoprecipitates derived from these strains(data not shown). However, the NOT3 protein wasable to be immunoprecipitated in the NOT2,not2<P59S, not2<W60C, not2<K134E andnot2<R165G strains. The CCR4–NOT complexcomponents NOT1, NOT2, NOT4, NOT5, CCR4and CAF40 were again all present in the immuno-precipitates derived from these strains (data notshown). One simple explanation for the effect ofthe not2 alleles on the CCR4–NOT complex is thatthey are affecting NOT3 association in the com-plex, which in turn is responsible for the pheno-types observed with not2-4 and not2<L9P. Thispossibility, however, does not seem to be the case,since we have previously demonstrated that a not3deletion neither affects the integrity of the CCR4–NOT complex nor results in phenotypes associatedwith not2-4 or not2<L9P.12

The NOT5 and NOT2 proteins of the CCR4–NOT complex have previously been described tohave a close protein–protein interaction.8 Toinvestigate the specific interaction between theNOT5 and NOT2 proteins in the mutant NOT2backgrounds, NOT5 immunoprecipitations wereperformed. Interestingly, the NOT2–NOT5 inter-action was stabilized in the strain containing thenot2<L9P mutation (Figure 3). This effect is notseen with the other NOT2 mutations including thenot2-4 allele (Figure 3). It should be noted thatwhile the NOT5 immunoprecipitations revealedmultiple species of NOT5 in the yeast extract noclear correlation between location of the mutationand the NOT5 banding pattern was observed.

Mutations in NOT2 specifically alter theprotein–protein interactions within the 1.9 MDaand 1.0 MDa CCR4–NOT complexes

While the above results indicate that the not2-4and not2<L9P proteins affect the overall integrityof the CCR4–NOT complex, the differing effects ofthese variants on LexA-NOT2 transactivation and

on the NOT5–NOT2 interaction suggested a morecomplex effect on the protein contacts within thecomplex. Moreover, immunoprecipitating CCR4–NOT components appears to co-immunoprecipi-tate proteins present only in the 1.0 MDa complexand may not be informative of the integrity of the1.9 MDa complex.12,13 To analyze these two com-plexes in more depth we examined the effect ofthe not2 alleles on the 1.9 MDa and 1.0 MDaCCR4–NOT complexes using Superose 6 gel chro-matography. We followed the migration of theCCR4, CAF1, NOT5, and CAF40 proteins in thesecomplexes. The location of NOT2 or NOT3 couldnot be determined since these proteins are difficultto detect following gel filtration chromatography.12

Analysis of the extracts from the strain contain-ing the not2-1 null mutation by gel filtrationchromatography revealed a complete loss of CCR4from the 1.9 MDa and 1.0 MDa CCR4–NOT com-plexes and the formation of a new complex withan average molecular mass of 0.7 MDa (Figure4(a)). Similar results were obtained for CAF1 andCAF40 (Figure 4(a)). This result confirms that theNOT2 protein is involved in protein–protein inter-actions that help to stabilize the CCR4–NOT com-plexes. No NOT5 protein was observed at anymolecular mass, suggesting that it had come outof the CCR4–NOT complexes and had beendegraded (data not shown). The 0.7 MDa complexprobably represents loss of the NOT2, NOT3,NOT4 and NOT5 proteins from the 1.0 MDa com-plex on the basis of the immunoprecipitation data.

In the not2-4 background CCR4, CAF1, andCAF40 presence in both the 1.9 MDa and 1.0 MDacomplexes was reduced (Figure 4(c)). Also, the1.9 MDa complex was reduced in size to 1.8 MDa,and the 1.0 MDa complex displayed a reducedmolecular mass of about 0.8 MDa. The NOT5 pro-tein disassociated from the 1.9 MDa complex andformed a complex at 0.2 MDa (fractions 16 and 17).

In contrast to the not2-4 allele, the not2<L9Pallele caused nearly complete loss of the 1.9 MDacomplex. As can be seen in Figure 4(d), CCR4,CAF1 and NOT5 proteins were not detected in theSuperose fractions where the 1.9 MDa complex

Table 5. The 1.9 MDa and 1.0 MDa CCR4–NOT complexes are altered by mutations in NOT2

1.9 MDa complex 1.0 MDa complex

Complex isintact

Mr

(MDa)NOT5 is in the

complexComplex is

intactMr

(MDa) Effects on the CCR4–NOT complexes

NOT2 þ 1.9 þ þ 1.0 Nonenot2-1 – – – – 0.7 Complete loss of 1.9 MDa and 1.0 MDa

destabilizednot2-4 þ (reduced) 1.8 – – 0.8 NOT5 and CAF1 fall out of 1.9 MDa complex

and 1.0 MDa destabilizednot2<L9P – – – – 0.7 Loss of 1.9 MDa complex and 1.0 MDa

destabilizednot2<P59S þ 1.9 þ (reduced) þ (reduced) 0.8 1.0 MDa complex distabilizednot2<W60C þ (reduced) 1.8 – þ 0.8 Loss of NOT5 from the 1.9 MDa complexnot2<K134E þ 1.9 þ þ (reduced) 1.0 1.9 MDa complex is destabilizednot2<R165G þ 1.8 þ þ (reduced) 1.0 1.9 MDa complex is nearly normal


should be detected. There was a small amount ofCAF40 detected at 1.7 MDa, suggesting that asmall remainder of the 1.9 MDa complex may bepresent in this strain. The 1.0 MDa complex wasalso reduced in size to 0.7 MDa. NOT5 was againdetected solely at 0.25 MDa. These results indicatethat although the not2<L9P allele caused similareffects as not2-4 on the structure of the CCR4–NOT complex as suggested by the immunoprecipi-tation analysis, gel permeation analysis indicatesthat not2<L9P displays much more severe effectson the complex. Moreover, the not2-4 allele, whichcauses more severe phenotypes than not2<L9P, islikely to be affecting other functions than simplythe structure of the CCR4–NOT complex.

The remaining not2 mutants displayed much lesssevere effects on the 1.9 MDa and 1.0 MDa com-plexes (Figure 5). The not2<W60C and not2<P59Salleles affected the association of NOT5 protein inthe 1.9 MDa complex and reduced the size of the1.0 MDa complex to 0.8 MDa (Figure 5(b) and (c)).The not2<R165G and not2<k134 alleles had littleeffect on the 1.9 MDa complex and some effects onthe concentration of components in the 1.0 MDacomplex (Figure 5(d) and (e)).

In summary (see Table 5 for a summary of thegel filtration analysis), Superose 6 fractionation ofcrude yeast extracts containing the mutant NOT2alleles in combination with the immunoprecipita-tion data indicates that NOT2 is required for theintegrity of the 1.9 MDa complex and the NOT2-5proteins in the 1.0 MDa complex. Also, the reten-tion of the NOT5 protein within the 1.9 MDacomplex appears to be specifically dependent onthe NOT2 protein. Overexpressing NOT5 alone,however, did not rescue not2-4 or not2<L9P pheno-types, suggesting that the phenotypes of thesemutations are due to their overall effects and notsimply the result of the loss of the NOT5 proteinfrom the complex (data not shown).

The NOT2–ADA2 interaction is abrogated bythe not2-4 and not2<R165G mutations

The above data indicate distinct functionaleffects of the not2 alleles, most of which may beattributed to disruptions of interactions within theCCR4–NOT complex. Previous work has shownthat NOT2 is able to interact with ADA2 in vitro,and that the not2-4 allele blocked the binding ofNOT2 to ADA2.9 To determine if the NOT2 mutantalleles described above could also alter the NOT2–ADA2 interaction, the LexA–NOT2 fusion proteinscontaining the not2 alleles were tested in a gluta-thione-S-transferase(GST)-ADA2 pull-down assay.

In agreement with the previous study,9 wild-typeLexA–NOT2 interacted with GST-ADA2 (Figure6). LexA-not2<L9P, LexA-not2<P59S, LexA-not2<W60C, and LexA-not2<K134E all retainedthe ability to physically interact with ADA2. Incontrast, the mutant proteins not2-4 andnot2<R165G when expressed as LexA fusionproteins did not bind ADA2 in the GST pull-down

assay. This indicates that the point mutation inLexA-not2<R165G is having the same effect onNOT2 interaction with ADA2 as was observedwith LexA-not2-4 protein interaction with ADA2.These effects on ADA2 contacts correlated withLexA–NOT2 ability to activate transcription(Table 4) and indicate that disruption of theCCR4–NOT complex is not a prerequisite to affect-ing NOT2–ADA2 interactions. We also found thatoverexpression of ADA2 could not complementthe not2-4 allele, confirming that not2-4 phenotypesresult from a combination of defects (data notshown).

Discussion

NOT2 is required for the integrity of theCCR4–NOT complexes

Six individual point mutations spanning theNOT2 gene were categorized by their ability tophysically and functionally alter NOT2 interactionwith components of and related to the CCR4–NOT complex. Our results show that NOT2 is anessential factor in the stability of the CCR4–NOTcomplex.

The importance of NOT2 to the integrity of thecomplex was supported by two experimentalobservations. First, the not2-1, not2-4 andnot2<L9P mutations result in the inability ofNOT2, NOT5, NOT3, and NOT4 to associate withthe complex as shown by immunoprecipitationanalysis. Second, gel filtration analysis confirmedcomplete loss of the 1.9 MDa complex in anot2<L9P background and the correspondingreduction in size of the 1.0 MDa complex to about0.7 MDa. This smaller complex appeared to containonly CCR4, CAF1, NOT1, CAF40, and CAF130, as

Figure 6. GST-ADA2 binding to NOT2 is disrupted bythe not2-4 and not2<R165 alleles. Yeast protein extracts(400 mg) from strain EGY188 containing each of theLexA–NOT2 fusions as indicated were incubated withGST-ADA2 or GST as described.32 Western analysis wasconducted with antibody directed against LexA.


evidenced from the immunoprecipitationanalysis, confirming that NOT2 is responsible forthe association of the NOT3-5 proteins with bothcomplexes.

The severe phenotypes observed with not2-4 hadpreviously been correlated with the inability tobind ADA2. However, not2<L9P displayed nearlythe same severity of phenotypes as the not2-4 allelealthough the not2<L9P protein was normal in itsability to contact ADA2. These results support theopposing model that the disruption of the NOT2-5proteins from the CCR4–NOT complexes result inthe severe phenotypes associated with these not2alleles. Disrupting NOT2–ADA2 interactionsresult in more mild effects (as observed in thenot2<R165G background). When loss of ADA2contacts was coupled with loss of CCR4–NOTcomplex integrity, as with the not2-4 allele, moresevere phenotypes were observed than with onlydisrupting the CCR4–NOT complex as found inthe strain carrying not2<L9P.

The N terminus of NOT2 affects contacts toNOT5, whereas both N and C-terminalsegments affect ADA2 contacts

Previous findings suggested that the relativelysmall, 26 kDa, NOT2 protein contains twodomains. The C terminus would be involved inCCR4–NOT specific interactions, and the N termi-nus would be involved in binding ADA2.9 Thefindings presented here indicate that this model isincomplete and that both N and C-terminal seg-ments of NOT2 contribute to its contacts to ADA2and that the N terminus of NOT2 is responsiblefor contacting NOT5 and stabilizing the NOT3-5protein association with the CCR4–NOT complexto the NOTs. This interpretation is supported byfour observations. First, the N-terminal L9 residueof NOT2 was shown to affect NOT5 binding;immunoprecipitation of the NOT5 protein revealedthat the not2<L9P mutation promotes an increasedassociation of NOT5 with NOT2 (Figure 3). Second,N-terminal NOT2 defects affected the integrity ofthe CCR4–NOT complexes. Third, the L9P, P59S,W60C, and R31G alterations in NOT2 all affectedNOT5 migration in the 1.9 MDa complex, indicat-ing that the N terminus of NOT2 stabilizes NOT5in this complex. Fourth, we showed that theLexA-not2<R165G protein has reduced transcrip-tional activation activity and like the LexA–not2-4protein displayed reduced binding to GST-ADA2(Figure 6). Clearly, both N and C-terminalsegments of NOT2 are involved in contactingADA2.

Although it has been proposed that the ADA2–NOT2 interaction is indirect,9 NOT2 was observedto bind ADA2 in the not2<L9P background whenthe 1.9 MDa CCR4–NOT complex was completelydestabilized (Figure 4(d)). Moreover, NOT2 couldnot be immunoprecipitated by CAF1 or NOT3 inthe not2<L9P strain, suggesting that it also wasnot present in the 0.7 MDa CCR4–NOT complex

observed in this strain. These observations suggestthat NOT2 can bind ADA2 in the absence of eitherthe 1.9 MDa or 1.0 MDa complexes. It is possiblethat the direct in vitro interaction between ADA2and NOT2 was not previously observed becauseNOT5, which is tightly associated with NOT2,12 isrequired for stabilizing the interaction to ADA2.Alternatively, NOT5 or another NOT protein maybe the direct contact to ADA2.

not2 mutations and transcriptional activation

The genetic screen that was utilized to detectseveral novel not2 alleles sought mutations thataugmented GAL4-NOT2 ability to activate GAL4-dependent reporter gene expression. When thesame mutations were analyzed as LexA fusionsfor activating LexA-dependent reporter geneexpression, four mutations did cause increasedtranscriptional activation, but one fusion, LexA-not2<R165G, displayed reduced activation func-tion. This difference between GAL4-not2<R165Gand LexA-not2<R165G may be attributed to theuse of different reporters, different effects ofNOT2<R165G on GAL4 and LexA DNA binding,or other inherent differences between the GAL4and LexA activation systems. We have observedsimilar differences for other types of GAL4 andLexA constructs, for example, fusions to certainregions of CAF1 or ADR1.

The cause for the increased activation ability ofthe other NOT2 variant proteins may not be fromany specific enhanced contacts to transcriptionalmachinery components. Instead, since all of thesemutations disrupted the CCR4–NOT complexes tosome extent with the apparent loss of NOT5 andNOT2 from either or both of the CCR4–NOT com-plexes, it is possible that the increased abundanceof the free form of NOT2 allows better contacts tobe made to the transcriptional machinery as com-pared to the NOT2 that is associated with theCCR4–NOT complexes.

It is also worth noting that NOT2 transcriptionalactivation ability does not correlate directly withits ability to bind ADA2. For example, not2mutations that augment activation ability do notdisplay altered binding to ADA2. NOT2, therefore,is playing several roles: (1) it is required for theassociation of NOT3-5 proteins in the CCR4–NOTcomplexes; (2) it apparently binds directly toNOT5 and stabilizes it in these complexes; (3) itbinds ADA2, which is partially related to itsability to enhance transcription; and (4) it appearsto make ADA2-independent contacts to enhancetranscriptional activation. Previous reports thatthe NOT proteins contact TFIID components8,10

suggest that NOT5 may be making these ADA2-independent contacts to TFIID components. NOT2may, therefore, be influencing NOT5 in theseinteractions.


Materials and Methods

Yeast strains and culture

Yeast strains are listed in Table 6 and were grown at30 8C on YEP or minimal medium lacking either uracilor both uracil and histidine. The carbon source was 4%(w/v) glucose unless otherwise indicated. b-Galactosi-dase assays were carried out as described.4 The reporterthat was used to measure transcriptional activation byLexA–NOT2 fusion proteins was the p34 reporter con-taining eight LexA binding sites upstream of thepromoter region.4

Generation and screening of NOT2 mutations

The complete yeast NOT2 open reading frame wasexpressed as a fusion protein to the GAL4 DNA bindingdomain (GAL4DBD) in the expression vector pPC97.This vector was used as a template for mutagenic PCRusing one PCR primer within the GAL4DBD codingregion and a downstream primer within the vector-derived ADH polyadenylation signal. PCR reactionconditions were as described.30 This PCR mixture wasco-transformed into MaV103 with pPC97 that had beenlinearized using Sal I and Not1. Transformants wereselected on leucine-deficient plates, followed by replicaplating to 10 mM 3-AT plates for selection of activationcompetent mutants. The phenotypes of mutants selectedin this manner were confirmed by growth on uracil-deficient plates and by blue color change on bufferedplates containing X-Gal. Plasmids were then rescuedfrom these strains, NOT2 coding sequences were charac-terized by sequencing, and these sequences were sub-cloned into pRS316 for testing of their ability tocomplement not2-1 phenotypes.

Sequencing and DNA constructions of the NOT2mutant alleles

DNA from pRS316, containing the mutant NOT2alleles, was sequenced using the following oligonucleo-tide primers.

NOT2up 50GCCGAATTCAGAGAAAAATTTGGTTT-AAAA 30; NOT2mid 50 ATTGGTGAATGATTCTGGTG30; NOT2down 50 TTTGGCCGTCGACCTGCAGACCTT-TCCCTATC GCC 30; NOT2C-term 50TAACAAAATATT-ACGATAATACATTC 30 and LexA202 50 GCTTCACCAT-TGAAGGGCTGGC 30.

The mutant NOT2 coding sequences were cloned asEco RI-Sal I PCR products into the Eco RI-Sal I sites ofpLex202-1.21,31 The LexA–NOT2 alleles were verified by

sequencing analysis. Expression of the LexA–NOT2fusion proteins containing the mutant NOT2 alleles wasconfirmed by Western analysis.

Immunoprecipitation

Immunoprecipitations were performed asdescribed1,4,8 with 2 mg of crude yeast extract preparedin Ip buffer (containing 20%(v/v) glycerol) using puri-fied anti-CAF1, anti-NOT3, and anti-NOT5 antibodies.Western analysis was conducted as described.7

Gel filtration chromatography

The Superose 6 HR10/30 column packed with pre-parative grade Superose 6 media1 was calibrated usingthe molecular mass standard mixture containing bluedextran (2000 kDa), thyroglobulin (669 kDa), and bovineserum albumin (66 kDa). The flow rate was 0.2 ml/minute, and 0.5 ml was collected per fraction.12 Thestandards eluted at 9.5 ml, 14.5 ml and 18 ml, respect-ively, and were used to calculate the molecular massesof the column fractions. Yeast extracts for gel filtrationanalysis were prepared as described.1

GST-pulldown assays

GST fusion proteins were expressed and bound to glu-tathione-agarose beads as described.32 The EGY188 yeaststrain, transformed with each of the LexA–NOT2 fusionproteins, was grown in 100 ml cultures overnight to midlog phase. Binding experiments were performed asdetailed.9 Specifically, yeast protein extracts (400 mg)were incubated with 1 mg of GST or GST-ADA2 on gluta-thione-agarose beads in 250 ml of binding buffer (20 mMTris (pH 7.4), 50 mM NaCl, 10 mM MgCl2, 5 mM EDTA,10% glycerol, 1 mM PMSF) at 4 8C for one hour with con-stant rocking. Samples were washed four times with0.7 ml of binding buffer, and separated on a SDS-10%PAGE gel. Western analysis was performed using anti-LexA antibody, and the ECL visualization system(Dupont/NEN).

Acknowledgments

We hank R. Young for providing yeast strains Z421and Z1446 and B. Lauze for aid in preparing thismanuscript. This research was supported by NIH grantGM41215 and HATCH 291 to C.L.D. This is scientificcontribution no. 2131 from the New HampshireAgricultural Experiment Station.

References

1. Liu, H.-Y., Badarinarayana, V., Audino, D. C.,Rappsilber, J., Mann, M. & Denis, C. L. (1998). TheNOT proteins are part of the CCR4 transcriptionalcomplex and affect gene expression both positivelyand negatively. EMBO J. 17, 1096–1106.

2. Chang, M., French-Corney, D., Fan, H., Klein, H.,Denis, C. L. & Jaehning, J. A. (1999). A complex con-taining RNA polymerase II, PAF1p, CDC73p,HPR1p, and CCR4p plays a role in protein kinase Csignaling. Mol. Cell Biol. 19, 1056–1067.

Table 6. Yeast strains used

Strain Genotype

KY803 MATa leu2-PET56 trp1-D1 ura3-5 gal2 gcn4-D1MY16 Isogenic to KY803 except not2-1MY22 Isogenic to KY803 except not2-4EGY188 MATa ura3 his3 trp1 LexA-LEU2MaV103 MATa ura3-52 leu2-3 112trp 1-901his3D200

SPAL10<URA3ade2-101 gal4D gal80D can1R

cyh2R GAL1<HIS3@LYS2 GAL1< lacZZ421 MATa ura3-52 his3D200 leu2-3 112 rbp2-3Z1446 MATa ura3-52 his3D200 leu2-3 112 rbp2-3, not2<P59S


3. Denis, C. L. (1984). Identification of new genesinvolved in the regulation of alcohol dehydrogenaseII. Genetics, 108, 833–844.

4. Draper, M. P., Liu, H.-Y., Nelsbach, A. H., Mosley,S. P. & Denis, C. L. (1994). CCR4 is a glucose-regulated transcription factor whose leucine-richrepeat binds several proteins important for placingCCR4 in its proper promoter context. Mol. Cell Biol.14, 4522–4531.

5. Liu, H.-Y., Toyn, J. H., Chiang, Y.-C., Draper, M. P.,Johnston, L. H. & Denis, C. L. (1997). DBF2, a cellcycle-regulated protein kinase, is physically andfunctionally associated with the CCR4transcriptional regulatory complex. EMBO J. 16,5289–5298.

6. Denis, C. L., Draper, M. P., Liu, H.-Y., Malvar, T.,Vallari, R. C. & Cook, W. J. (1994). The yeastCCR4 protein is neither regulated by norassociated with the SPT6 and SPT10 proteins andforms a functionally distinct complex from that ofthe NF/SWI transcription factors. Genetics, 138,1005–1013.

7. Draper, M. P., Salvadore, C. & Denis, C. L. (1995).Identification of a mouse protein, whose homolog inSaccharomyces cerevisiae is a component of the CCR4transcriptional regulating complex. Mol. Cell Biol. 15,3487–3495.

8. Badarinarayana, V., Chiang, Y.-C. & Denis, C. L.(2000). Functional interaction of CCR4–NOTproteins with TATAA-binding protein (TBP) andits associated factors in yeast. Genetics, 155,1045–1054.

9. Benson, J. D., Benson, M., Howley, P. M. & Struhl, K.(1998). Association of distinct yeast NOT2 functionaldomains with components of GCN5 histoneacetylase and CCR4 transcriptional regulatory com-plexes. EMBO J. 17, 6714–6722.

10. Lemaire, M. & Collart, M. (2000). The TATA-bindingprotein associated factor yTafII19p functionally inter-acts with components of the global transcriptionalregulator CCR4–NOT complex and physically inter-acts with the NOT5 subunit. J. Biol. Chem. 275,26925–26934.

11. Liu, H.-Y., Chiang, Y.-C., Pan, J., Salvadore, C., Chen,J., Audino, D. C. et al. (2001). Characterization ofCAF4 and CAF16 reveal a functional connectionbetween the CCR4–NOT complex and a subset ofSRB proteins of the RNA polymerase II holoenzyme.J. Biol. Chem. 276, 7541–7548.

12. Bai, Y., Salvadore, C., Chiang, Y.-C., Collart, M. A.,Liu, H.-Y. & Denis, C. L. (1999). The CCR4 andCAF1 proteins of the CCR4–NOT complex are physi-cally and functionally separated from NOT2, NOT4and NOT5. Mol. Cell Biol. 19, 6642–6651.

13. Chen, J., Chiang, Y.-C., Rappsilber, J., Russell, P.,Mann, M. & Denis, C. L. (2001). Purification andcharacterization of the 1.0 MDa CCR4–NOTcomplex identifies two novel components of thecomplex. J. Mol. Biol. 314, 683–694.

14. Hata, H., Mitsui, H., Liu, H., Bai, Y., Denis, C. L.,Shimizu, Y. & Sakai, A. (1998). DHH1p, a putativeRNA helicase, associates with the general transcrip-tion factors POP2p and CCR4p from Saccharomycescerevisiae. Genetics, 148, 571–579.

15. Komarnitsky, S. I., Chiang, Y.-C., Luca, F. C., Chen, J.,Toyn, J. H., Winey, M. et al. (1998). DBF2 proteinKinase binds to and acts through the cell cycle-regulated MOB1 protein. Mol. Cell Biol. 18,2100–2107.

16. Chen, J., Chiang, Y.-C. & Denis, C. L. (2002). CCR4, a30-50 poly(A) RNA and ssDNA exonuclease, is thecatalytic component of the cytoplasmic deadenylase.EMBO J. 21, 1414–1426.

17. Daugeron, M.-C., Mauxion, F. & Seraphin, B. (2001).The yeast POP2 gene encodes a nuclease involvedin mRNA deadenylation. Nucl. Acids Res. 29,2448–2455.

18. Tucker, M., Staples, R. R., Valencia-Sanchez, M. A.,Muhlrad, D. & Parker, R. (2002). CCR4p is thecatalytic sub-unit of Ccr4p/Pop2p/Notp mRNAdeadenylase ecomplex in Saccharomyces cerevisiae.EMBO J. 21, 1427–1436.

19. Tucker, M., Valencia-Sanchez, M. A., Staples, R.,Chen, J., Denis, C. L. & Parker, R. (2001). The tran-scription factor associated Ccr4 and Caf1 proteinsare components of the major cytoplasmic mRNAdeadenylase in Saccharomyces cerevisiae. Cell, 104,377–386.

20. Collart, M. A. & Struhl, K. (1993). CDC39, anessential nuclear protein that negatively regulatestranscription and differentially affects the con-stitutive and inducible HIS3 promoters. EMBO J. 12,177–186.

21. Collart, M. A. & Struhl, K. (1994). NOT1 (CDC39),NOT2 (CDC36), NOT3, and NOT4 encode a global-negative regulator of transcription that differentiallyaffects TATA-element utilization. Genes Dev. 8,525–537.

22. Cade, M. R. & Errede, B. (1994). MOT2 encodes anegative regulator of gene expression that affectsbasal expression of pheromone-responsive genes inSaccharomyces cerevisiae. Mol. Cell Biol. 14, 3139–3149.

23. Collart, M. A. (1996). The NOT, SPT3, and MOT1genes functionally interact to regulate transcriptionat core promoters. Mol. Cell Biol. 16, 6668–6676.

24. Irie, K., Yamaguchi, K., Kawase, K. & Matsumoto, K.(1994). The yeast MOT2 gene encodes a putative zincfinger protein that serves as a global negativeregulator affecting expression of several categoriesof genes, including mating-pheromone responsivegenes. Mol. Cell Biol. 14, 3150–3157.

25. Albert, T. K., Hanzawa, H., Legtenberg, Y. I. A., deRuwe, M. J., van den Heuvel, F. A. J., Collart, M. A.et al. (2002). Identification of a ubiquitin-proteinligase subunit within the CCR4–NOT transcriptionrepressor complex. EMBO J. 21, 355–364.

26. Vidal, M., Brachmann, R. K., Fattaey, A., Harlow, E.& Boeke, J. D. (1996). Reverse two-hybrid and one-hybrid systems to detect dissociation of protein–protein and DNA–protein interactions. Proc. NatlAcad. Sci. USA, 93, 10315–10320.

27. Hope, I. A. & Struhl, K. (1986). Functional dissectionof a eukaryotic transcriptional activator protein,GCN4 of yeast. Cell, 46, 885–894.

28. Lee, T. I., Wyrick, J. J., Koh, S. S., Jennings, E. G.,Gadbois, E. L. & Young, R. A. (1998). Interplay ofpositive and negative regulators in transcriptioninitiation by RNA polymerase II holoenzyme. Mol.Cell Biol. 18, 4455–4462.

29. Denis, C. L., Chiang, Y.-C., Cui, Y. & Chen, J. (2001).Genetic evidence supports a role for the yeastCCR4–NOT complex in transcriptional elongation.Genetics, 158, 627–634.

30. Yasugi, T., Vidal, M., Sakai, H., Howley, P. M. &Benson, J. D. (1997). Two classes of humanpapillomavirus type 16 E1 mutants suggest pleio-tropic conformational constraints affecting E1


multimerization, E2 interaction, and interaction withcellular proteins. J. Virol. 71, 5942–5951.

31. Brent, R. & Ptashne, M. (1985). A eukaryotic tran-scriptional activator bearing the DNA specificity ofa prokaryotic repressor. Cell, 43, 729–736.

32. Chiang, Y.-C., Komarnitsky, P., Chase, D. & Denis,C. L. (1996). ADR1 activation domains contact thehistone acetyltransferase GCN5 and the core tran-scriptional factor TFIIB. J. Biol. Chem. 271,32359–32365.

Edited by S. Reed

(Received 1 March 2002; received in revised form 18 June 2002; accepted 9 July 2002)


Characterization of Mutations in NOT2 Indicates that it Plays an Important Role in Maintaining the...

Documents

Transcript of Characterization of Mutations in NOT2 Indicates that it Plays an Important Role in Maintaining the...